Back
Exploring Perturbed Physics Ensembles in a Regional Climate Model
Journal article

Exploring Perturbed Physics Ensembles in a Regional Climate Model

  • Bellprat, Omar Institute for Atmospheric and Climate Science, ETH Zurich, Zurich, Switzerland
  • Kotlarski, Sven Institute for Atmospheric and Climate Science, ETH Zurich, Zurich, Switzerland
  • Lüthi, Daniel Institute for Atmospheric and Climate Science, ETH Zurich, Zurich, Switzerland
  • Schär, Christoph Institute for Atmospheric and Climate Science, ETH Zurich, Zurich, Switzerland
  • 2012-7-9
Published in:
  • Journal of Climate. - American Meteorological Society. - 2012, vol. 25, no. 13, p. 4582-4599
Atmospheric Science
English Abstract
Perturbed physics ensembles (PPEs) have been widely used to assess climate model uncertainties and have provided new estimates of climate sensitivity and parametric uncertainty in state-of-the-art climate models. So far, mainly global climate models were used to generate PPEs, and little work has been conducted with regional climate models. This paper discusses the parameter uncertainty in two PPEs of a regional climate model driven by reanalysis data for the present climate over Europe. The uncertainty is evaluated for the variables of 2-m temperature, precipitation, and total cloud cover, with a focus on the annual cycle, interannual variability, and selected extremes.
The authors show that the simulated spread of the PPEs encompasses the observations at a regional scale in terms of the annual cycle and the interannual variability, provided observational uncertainty is taken into account. To rank the PPEs a new skill metric is proposed, which takes into account observational uncertainty and natural variability. The metric is a generalization of the climate prediction index (CPI) and is compared to metrics used in other studies. The consideration of observational uncertainty is particularly important for total cloud cover and reveals that current observations do not allow for a systematic evaluation of high precipitation intensities over the entire European domain.
The skill framework is additionally used to identify important model parameters, which are of interest for an objective model calibration.
Language
  • English
Open access status
hybrid
Identifiers
Persistent URL
https://sonar.ch/global/documents/107998
Statistics
Document views: 19 File downloads: